95th ESA Annual Meeting (August 1 -- 6, 2010)

COS 70-7 - Using occupancy models to investigate large-scale stem-boring insect herbivory dynamics in a primary successional landscape

Wednesday, August 4, 2010: 3:40 PM
410, David L Lawrence Convention Center
Christian Che-Castaldo, Biology Department, University of Maryland, College Park, MD, Evan H. Campbell Grant, Patuxent Wildlife Research Center, US Geological Survey, William F. Fagan, Department of Biology, University of Maryland, College Park, MD, Charlie Crisafulli, Mount St. Helens National Volcanic Monument, US Forest Service and John G. Bishop, Biology, Washington State University, Vancouver, WA
Background/Question/Methods

Recent studies of primary succession on Mt. St. Helens have focused on the hypothesis that insect herbivory is a major factor influencing the pace and pattern of early primary succession, not a minor mechanism as implied in nearly all work to date. Evaluating this hypothesis necessitates collecting data on plant-insect herbivore interactions in primary successional landscapes. We have collected a large-scale observational dataset in 2009 of herbivory patterns by two stem boring insects (the weevil Cryptorynchus lapathi and the sesiid moth Paranthrene robiniae) that attack Sitka willow (Salix sitchensis) on the Mt. St. Helens Pumice Plain, a 15 km2 landscape of rock and ash formed and sterilized by pyroclastic flows during the 1980 eruption. Willow is a key successional species because it is the first Pumice Plain plant to provide three-dimensional structure and microclimates necessary for bird and small mammal colonists. We use a combined occupancy modeling and information theoretic approach to estimate the unbiased proportion of sites (stems) occupied by stem-boring insect larvae and investigate possible hypotheses explaining variation in herbivory patterns at the landscape level. 
Results/Conclusions

We treat stems as sites and model stem-boring insect herbivory as occupancy of these sites using a single-species, single-season occupancy model. We compete three hypotheses regarding how plant water stress, plant gender, and/or stem reproductive status explain variation in herbivore occupancy patterns by incorporating these factors as stem-level covariates. We evaluate support for the inclusion of the covariates as important predictors using an information-theoretic approach based on AIC and model weights. Based on this analysis, the top models include 83% of the model weight. These top models feature an interactive effect of reproductive status and habitat while gender is included as either a main effect or in an interaction. We found that reproductive stems are attacked more frequently by stem-borers than non-reproductive stems, and this effect is more pronounced in upland habitat. Male stems are attacked less frequently than female stems regardless of habitat and reproductive status. These results suggest that gender, water stress, and reproductive status all affect large-scale dynamics of stem-boring herbivores, possibly via differences in willow stem nutrient and/or secondary metabolite content.